| Home | Publications | Proceedings |
Objectives: In recent years, considerable progress has been made toward accurate, noninvasive determination of long bone strength. Typically, bone mineral mass or density is measured at specific skeletal sites using densitometry or computed tomography, but there has been considerable attention to the need to also assess distribution of bone tissue for optimal prediction of strength. Debate continues, however, regarding which bone strength parameter, skeletal site, and assessment technique is most appropriate for the prediction of fracture risk in a clinical setting. Our objectives were to: (i) determine the parameter(s) which best predict whole bone fracture and (ii) determine the best skeletal site at which to make measurements.
Clinical Relevance: Although it is well accepted that bone mass is not the only determinant of bone strength, we were only able to demonstrate a small advantage in prediction of whole bone failure by inclusion of geometric properties in our assessment. Further, we found that the midshaft assessments were more accurate in predicting whole bone failure than measurements taken after the fact at the site of fracture. These findings indicate that standard bone densitometry is robust as a clinical assessment technique. Future work needs to investigate what other properties may be measured relevant to bone failure, such as material properties, tissue-level defects, anisotropy, and bone shape.
Methods: Thirteen pairs of cadaver femora were scanned using dual energy x-ray absorptiometry with radiographic images obtained concurrently. From these data, the following bone strength parameters were calculated along the length of the femoral shaft: bone mineral content (BMC); cross-sectional area, polar moment of inertia and polar section modulus using radiographic measurements and a concentric annulus model; cross-sectional area, cross-sectional moment of inertia, and polar section modulus obtained directly from raw absorptiometry data using custom written software; and a strength indicator (SI) parameter based on beam theory which incorporates both BMC and polar section modulus. The left femurs were subsequently tested to failure in torsion. Correlation analysis was performed between torsional failure load and each parameter measured at the midshaft and fracture sites.
Results: All parameters showed the same basic pattern of variation along the femoral shaft. Geometric and structural properties were maximized in the proximal region and minimized in the distal region. The correlation between failure load and parameters measured at the fracture site was generally low. The correlation between failure load and parameters measured at midshaft were considerably higher, with parameters that incorporated both amount and distribution of mineral showing the strongest association with failure load.
Conclusions: In most of the specimens, bone failure did not occur in the region where both structural and geometric properties were minimized, and data obtained at the midshaft were more accurate in predicting fracture than data obtained at the actual site of fracture. Parameters that incorporated both mass and distribution of bone mineral had the strongest predictive power; however, BMC alone provided a reasonable prediction of fracture risk. With currently available assessment technology, the added time and effort needed to calculate cross-sectional properties may not be justified for a relatively small gain in predictive power.
Acknowledgments: VA Medical Research Service (Regional Advisory Group); Palo Alto Institute for Research and Education.